An expert system for preliminary landfill design in developing countries

The application of expert systems technology in the domain of sanitary landfill design for developing countries is particularly appropriate in order to preserve and disseminate valuable experience efficiently at reasonable costs. Despite the simplicity in the design and operational concepts of landfilling works, there is a conspicuous scarcity in available expertise for its implementation. The objective of the research is to develop an expert sytem prototype, the Landfill Design Advisor (LDA), for the preliminary phase in the design of sanitary landfills, with particular considerations to the peculiar characteristics of developing countries. It is an attempt to harness expertise in an important domain and exploit the potential of the latest in computing technology to produce a user-friendly system for the non-computer specialist. The two major issues addressed by the research are knowledge acquisition and prototype building. The knowledge acquisition process is widely acknowledged by leading artificial intelligence researchers to be the single most critical obstacle to the development of effective expert systems. The knowledge acquisition process for the LDA adopted a methodology where expertise is acquired from a variety of sources using suitable techniques such as text analysis, unstructured interviews and observation of landfilling works. The prototype building process for the LDA adopted a modular approach using the incremental rapid prototyping methodology. An exploratory test prototype was first developed using a rule-based expert system shell, Xi-Plus, to demonstrate the feasibility of this research. An alternative shell, Kappa-PC, was then considered and used for developing the main LDA system. It was preferred because of its object-oriented approach which combines the two main knowledge representation techniques for expert systems, i. e. production rules and objects (or frames).